Environmental causality calibration: Advancing WLAN RF fingerprinting for precise indoor localization.

Area of Science:

• Electrical Engineering
• Computer Science
• Signal Processing

Background:

• Wireless Local Area Network (WLAN) Radio Frequency (RF) fingerprinting is a promising indoor positioning technology.
• Indoor environmental factors, especially relative humidity, significantly degrade positioning accuracy by affecting RF signal propagation.
• Existing research lacks comprehensive analysis of humidity's impact on WLAN RF signals and positioning.

Purpose of the Study:

• To propose and evaluate an Adaptive expansion fingerprint database (AeFd) model.
• To mitigate the impact of indoor relative humidity variations on WLAN RF fingerprinting accuracy.
• To enhance the stability and performance of indoor positioning systems.

Main Methods:

• Developed an Adaptive expansion fingerprint database (AeFd) model using a regression learning algorithm.
• Designed a relationship model to describe fingerprint database interactions under varying relative humidity.
• Integrated the AeFd model with the K-Nearest Neighbors (KNN) algorithm for experimental validation.

Main Results:

• The AeFd model effectively expands the fingerprint database across different relative humidity levels.
• Experimental results demonstrated a 5% performance improvement over 10 days and an 8% improvement over 10 months when using AeFd with KNN.
• The proposed model significantly enhances positioning performance and system stability.

Conclusions:

• The Adaptive expansion fingerprint database (AeFd) model successfully addresses the challenge of humidity-induced accuracy degradation in WLAN RF fingerprinting.
• The AeFd model offers a viable solution for dynamic adaptation to environmental changes, improving the robustness of indoor positioning systems.
• This research contributes to more reliable and stable indoor positioning through effective environmental factor mitigation.